The value here can change depending on the type that PAGE_SIZE
has on a given architecture. To avoid the ensuing signed and
unsigned division conversions, we shift instead using PAGE_SHIFT
Signed-off-by: Daniel Rosenberg <drosen@google.com>
Bug:
35257858
Change-Id: I132cae93abea39390c3f0f91a4b2e026e97ed4c7
int start_idx, end_idx;
struct squashfs_sb_info *msblk = req->sb->s_fs_info;
- start_idx = (idx * msblk->devblksize - req->offset) / PAGE_CACHE_SIZE;
- end_idx = ((idx + 1) * msblk->devblksize - req->offset + 1) / PAGE_CACHE_SIZE;
+ start_idx = (idx * msblk->devblksize - req->offset) >> PAGE_SHIFT;
+ end_idx = ((idx + 1) * msblk->devblksize - req->offset + 1) >> PAGE_SHIFT;
if (start_idx >= req->output->pages)
return 1;
if (start_idx < 0)